[rfc-i] Current Cacheing behavior for current XML2RFC

Jim Schaad ietf at augustcellars.com
Tue Jan 7 17:46:14 PST 2014


There are a number of issues of caching of references on a local machine, I
am not sure what the current behavior is for the server but it could have
some of the same types of interesting problems as well.

 

1.        The path that is used for caching of references on the local
machine is coded into the software as ['/var/cache/xml2rfc',
'~/.cache/xml2rfc'].  A command line option exists to insert a new directory
at the front of this set of paths.  I am not sure how the first directory
was arrived at, the second one I recognize as being per user specific.   It
is not clear that having an absolute path encoded into the system makes
sense for Linux, it is really a potential problem for Windows unless we are
going to tell people where it is going to exist before we start.  

2.       The current cache can be filled and emptied, however it is never
checked and refreshed in any sense.  I am going to assume that there is a
script on the servers that clear the cache between any pair of invocations
of the script (or that something similar is done to get things refreshed).

 

If I am really the only person that is running Xml2rfc locally on a Windows
system, then there is no real reason the try and fix the first one so that
the paths are going to make more sense.  I have figure out where the caches
are and how to play with them so it falls into the no-harm no-foul rule.

 

I would however like to see the second issue fixed in one of a couple of
different ways.  There are some easy ways and some harder ways to fix the
problem that would have different behaviors on the user experience.

 

a.       Just delete any item that is in the cache longer than a specific
amount of time and force it to be re-downloaded.  This means that documents
which are no longer current will disappear, however if you go off line you
might not be able to use that cached file at a later date.

b.      If an item is over than a specific amount of time, then try and down
load it again.  However if you fail then just leave the current item in the
cache.  This would mean that updates would be found, however items that
disappear would never vanish from the cache unless the user uses the clear
cache option on the command line (or deletes the contents of the cache).

c.       Something else that might make sense

 

Jim

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.rfc-editor.org/pipermail/rfc-interest/attachments/20140107/c9c3dae7/attachment.htm>


More information about the rfc-interest mailing list